Statistics from Altmetric.com
We read with interest the letter published by Vu et al1 which investigated the risk of age related macular degeneration (AMD) and its association with the dietary carotenoids, lutein and zeaxanthin (LZ), stratified by linoleic acid intake. Vu et al reported a marked increase in the risk of both early and late AMD among people who consumed greater than the median intake of linoleic acid and higher dietary intakes of LZ.
We have a number of concerns in relation to the authors’ letter and their conclusions. The letter used cross sectional data based on photographic macular assessments of 71.9% of their sample of 2448 people, who attended follow up examinations. The authors also included 212 people who did not have photographic macular assessment (10.8% of those with dietary assessments). This is one area of concern. The dietary assessment method (food frequency questionnaire, FFQ) was not conducted at baseline, which only allows measurements of association from the follow up examination. Owing to the cross sectional nature of the data, it is plausible and even likely that participants with known signs of early macular degeneration or associated visual changes may have increased their dietary antioxidant intakes (indication bias)—for example, after being told about their signs at the first examination or at other times. This bias may have occurred in particular among those consuming higher linoleic acid diets as higher intakes of linoleic acid have been reported to increase the risk of AMD.2 An excellent example of this indication bias can be seen in the finding of a significantly increased risk of poor night vision associated with increased consumption of carrots.3
The letter also states that a possible protection existed with high LZ intake on AMD among those with low levels of linoleic acid intake. We could, however, not see any data in the results or tables to support this statement.
We thought that these findings needed to be confirmed in other study populations. Given our concerns about the cross sectional data design we explored this association with the incidence of AMD in the Blue Mountains Eye Study cohort. Baseline data were collected in 1992–4 from 3654 residents of the Blue Mountains aged 49 years and over. Eye examinations were conducted on residents at baseline and at follow up using retinal photographs and the Wisconsin AMD grading system.4 Dietary data were assessed at baseline (n = 2900) and follow up examinations using a 145 item food frequency questionnaire (FFQ).5 Of the participants examined at either or both the 5 year and 10 year examinations, 2454 had retinal photographs available for the assessment of age related maculopathy (ARM) lesions. Of the 2454 participants 2083 had complete FFQ data, including 818 supplement users. We used the Willett method6 to energy adjust the linoleic and LZ data and investigated those with less than and greater than median intake for linoleic acid (median = 6.6 g) and 1 standard deviation increases of LZ (mean intake of energy adjusted LZ intake was 819 μg, with an SD of 475 μg), using a multivariate adjusted discrete logistic model to assess factors associated with 10 year incident AMD.
We found no association with energy adjusted LZ intake and the incidence of early, late, or any AMD, whether or not this was stratified by linoleic acid intake (table 1). Given that our median linoleic acid intake was less than the median used by Vu et al1 (6.6 g verses 7.2 g) we also stratified the data by the highest tertile of linoleic acid intakes (cut point 8.5 g) and also found no association between LZ and incident AMD (data not shown).
While the examination of cross sectional data to investigate associations with disease may be useful, conclusions drawn from such data need to be made with care in the light of other known literature. Other data have demonstrated a potential protective effect from foods containing LZ7 and foods and supplements rich in LZ are known to increase (protective) macular pigment density.8 In addition, dietary guidelines9 include recommendations to increase vegetables and fruit that are rich in LZ (for example, broccoli, green beans, silverbeet, brussel sprouts, oranges).10 The authors’ conclusions, based on their very limited data, are non-intuitive and we are concerned that they are potentially misleading. Care needs to be taken to continue to encourage the vegetable and fruit intakes of populations.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.